The paper introduces Fuzzy Identity-Based Encryption (Fuzzy IBE), a new type of Identity-Based Encryption (IBE) scheme that allows for error-tolerance between the identities of a private key and the public key used to encrypt a ciphertext. In Fuzzy IBE, identities are viewed as sets of descriptive attributes, and a private key for an identity $\omega$ can decrypt a ciphertext encrypted with an identity $\omega'$ if and only if $\omega$ and $\omega'$ are close to each other in terms of set overlap. This property enables the use of biometric inputs as identities, which inherently have some noise, and also supports attribute-based encryption. The authors present two constructions of Fuzzy IBE schemes. The first construction uses groups with an efficiently computable bilinear map and assumes the hardness of the Computational Diffie-Hellman problem. The second construction, designed for large universes where attributes are defined by arbitrary strings, also reduces to the hardness of the Decisional Bilinear Diffie-Hellman assumption. The paper discusses the security of the schemes under the Selective-ID model and proves their security without relying on random oracles. The authors also explore the efficiency and key sizes of the schemes, showing that the public parameters, private keys, and ciphertexts grow linearly with the number of attributes or the size of the identity being encrypted. Finally, the paper concludes by discussing potential future work, including the possibility of creating Fuzzy IBE schemes with multiple authorities and using different distance metrics between identities.The paper introduces Fuzzy Identity-Based Encryption (Fuzzy IBE), a new type of Identity-Based Encryption (IBE) scheme that allows for error-tolerance between the identities of a private key and the public key used to encrypt a ciphertext. In Fuzzy IBE, identities are viewed as sets of descriptive attributes, and a private key for an identity $\omega$ can decrypt a ciphertext encrypted with an identity $\omega'$ if and only if $\omega$ and $\omega'$ are close to each other in terms of set overlap. This property enables the use of biometric inputs as identities, which inherently have some noise, and also supports attribute-based encryption. The authors present two constructions of Fuzzy IBE schemes. The first construction uses groups with an efficiently computable bilinear map and assumes the hardness of the Computational Diffie-Hellman problem. The second construction, designed for large universes where attributes are defined by arbitrary strings, also reduces to the hardness of the Decisional Bilinear Diffie-Hellman assumption. The paper discusses the security of the schemes under the Selective-ID model and proves their security without relying on random oracles. The authors also explore the efficiency and key sizes of the schemes, showing that the public parameters, private keys, and ciphertexts grow linearly with the number of attributes or the size of the identity being encrypted. Finally, the paper concludes by discussing potential future work, including the possibility of creating Fuzzy IBE schemes with multiple authorities and using different distance metrics between identities.